George Kenwright

Creative producer specialising in AI production. Google DeepMind.

[JY @agenticjay]

@GeokenAI's Argon Pagoda but make it @Aprilzz423’s colorfigures SVG skill with Gemini 4 Argon.

Single shot, simple prompt. Skill + Reference video.

Argon's ability to understand references and recreate with a simple loop are mind blowing. The consistency every time gets me.

[George Kenwright @GeokenAI]

So cool!

[Boris Starkov @ktoya_me]

Introducing https://t.co/427WVklG7W

A massive isometric pixel map of London, based on amazing NYC demo by @_coenen from 9 months ago

My agent created the dataset, finetuned qwen image LoRA adapter, did a multiday inference run and built an efficient tile viewer!

[AshutoshShrivastava @ai_for_success]

11 hours with Gemini 4 Argon, Google literally cooked with Gemini Argon. 🔥
it's insane for building 3D stuff. A full Asuka-period Japanese pagoda temple, built in Blender and rendered in Cycles. The level of precision in everything is just crazy.

[George Kenwright @GeokenAI]

@Sachi_0298 @GoogleAI @GoogleDeepMind @antigravity Loop 1 got me me 90% of the way. The remaining 10% took another ~30 loops. That might seem unbalanced, but that final 10% is the difficult/important bit: identifying and correcting subtle problems in lighting and design.

[Google AI @GoogleAI]

See a fully-explorable 3D world build itself in minutes ⛩️

Watch @GoogleDeepMind's @GeokenAI explain how Gemini 4 Argon acts as a virtual builder in @antigravity to autonomously create a stunning, multi-tiered pagoda sanctuary.

[0xbitcat @_0xbitcat]

终于完成了，复刻了@GoogleDeepMind发的@GeokenAI 利用Gemini 4 Argon制作的演示视频。昨天看完视频之后非常激动，立马订阅了ultra然后上手尝试，选择了中国传统建筑作为主题。根据Argon自己的统计，花费大约12小时，累计token吞吐量大约100M。

[Miao Yang @WaterrrMiao]

the part of argon i love most is the critique loop. one agent literally looks at the render and goes "too blocky" before it gets fixed

so that's the turning point of this video: wooden toy pagoda → photoreal temple, same shot

congrats to the gemini team 👏 made in hyperframes studio with one prompt

[Ben Cobley @cobley_ben]

3D-to-2D animation with Argon in Antigravity

using the amazing pagoda video mp4 as the input from @GeokenAI + 12fps makes it feel more anime :)

[George Kenwright @GeokenAI]

@ai_for_success No way. Can't wait to see the results!

[George Kenwright @GeokenAI]

In the future everyone will have their own photorealistic 3D pagoda

[AshutoshShrivastava @ai_for_success]

Gemini 4 Argon in Antigravity is an absolute beast 🔥

Inspired by yesterday’s Gemini 4 Argon demo by George Kenwright, I wanted to recreate a similarly high-quality and accurate 3D structure.

I fed the demo video to Antigravity and asked it to reverse-engineer the skill behind it, using the visible workflow as a reference. I then upgraded my existing skill based on the process used in the official demo.

4 hours in, and Antigravity is absolutely cooking. Let's see how long it takes, but so far, it's looking really good!

More details soon :)

[George Kenwright @GeokenAI]

@ai_for_success Great to see it can be recreated! The shots at 0:21 are amazing. Image generations or renders?

[George Kenwright @GeokenAI]

https://t.co/hGyKqDopdJ

[Google AI @GoogleAI]

🎉 Today, we’re making Gemini 2.5 Deep Think available to Google AI Ultra subscribers in the @GeminiApp

Gemini 2.5 Deep Think is a variation of the model that achieved the gold-medal standard at this year’s International Mathematical Olympiad (IMO). Deep Think uses parallel thinking and reinforcement learning techniques to deliver more detailed and thoughtful responses.

To demonstrate this, we prompted the model to generate a voxel art image of a pagoda using a single prompt. As you can see, compared to 2.5 Pro, Deep Think adds more and more detail to every aspect of the scene. From the architecture of the pagoda, to the fidelity of the trees, to the diversity of the ground plane, every aspect of the prompt gets handled more deeply.

[George Kenwright @GeokenAI]

@gmartidesign No manual adjustments to camera angle, this is straight out of the model. The point of this meta-demo is that Argon can do a good job without much steering. In different circumstances I'd give the model the exact angle and focal length for accuracy.

[George Kenwright @GeokenAI]

This is the real me, but the background is courtesy of Gemini 4 Argon.

I showed Argon the two green-screen stills below and asked it to 1) build me a room in Blender, 2) recreate the camera angles, 3) use Nano Banana to 'skin' the Blender background to match the foreground, and 4) key it back into the original video.

The result: an accurate background across both angles.

[Google DeepMind @GoogleDeepMind]

Gemini 4 Argon is now globally available.

Built to solve your hardest problems from start to finish, the model brings frontier capabilities across coding, enterprise workflows, and cyber defense.

[George Kenwright @GeokenAI]

@sebuzdugan I can't vouch for exactly how accurate the angle is... Tbh with this kind of green screen compositing you can get away with it not being perfect -- it just needs to look right! It would be interesting to test it more precisely.

[AshutoshShrivastava @ai_for_success]

Everyone was in love with this Gemini 4 Argon demo but the coolest part was @GeokenAI used Argon to build a room in Blender, recreate the camera angles, texture the background with Nano Banana, and composite it into the original green-screen video.

[JY @agenticjay]

Today we launched Gemini 4 Argon. Charts can tell you how it performs, but what really matters to me is it's just a joy to build with. Read about it in the blog, check out all our demos, and give it a try yourself! So proud of our team.

15 years building products has brought me to realize that the "perfect job" isn’t a single destination. It’s a three-sided puzzle of: (1) what you’re good at, (2) what makes you happy, and (3) what pays the bills.

With today’s launch, that puzzle led me to my recently new role within Google leading Demo Lab for Gemini models. With an incredible team around me, we shipped dozens of demos to share just what Argon can be capable of and spark imagination on the timeline.

I feel privileged to work alongside some of the most brilliant marketers, researchers, and creative technologists at Google, pushing our newest models in the most creative and technically impressive ways.

We build things that make people pause and think, "Wait, if it can do that, what could I build with it?" and while 3d demos alone won’t solve the world’s toughest challenges, my hope is they inspire the people that will to use our models to try. Billions of people around the world use Gemini models in their own lives to solve real-world problems. That’s impact.

To celebrate today’s launch of Gemini 4 Argon, here is one more demo. A 3D Rubik’s Cube I built in three.js because apparently my idea of unwinding after a massive product launch is arguing with quaternions at 12 AM. And yes, I just learned what a quaternion was today from Argon.

Whether it’s a 3×3 cube, the cure for a rare disease, or life’s puzzle of a career, know there’s no real magic to finding the solution, just a series of turns, one face at a time.

[genevieveh@ @genevieve__h]

I gave Gemini 4 Argon our Drive folder with launch videos, our brand + motion guidelines, and the challenge to make a cool trailer for itself. I love Argon's creativity - I did not write any of the text. Built with HyperFrames, music by Lyria (also prompted by Argon) 🚀

[Sundar Pichai @sundarpichai]

Generating complex, multi-tiered architecture is a great test for frontier models.

Using multi-agent visual critique loops in Google Antigravity, Argon autonomously created this highly detailed Pagoda and surrounding grounds from scratch in Blender.

[George Kenwright @GeokenAI]

@korymath Thanks @korymath !

[genevieveh@ @genevieve__h]

Gemini 4 Argon is here!!! I can't wait to see what people build with it, and I have so much to say/post/share but to start:

I used Argon to turn pictures of my home office desk into a playable racing game with Blender MCP and Three.js. How-to below: 🧵